Yesterday's Gallup Poll Had GOP Advantage - Once Again
Yesterday, Gallup released a new poll for USA Today and CNN, which showed a tightened race in both of their registered voter (RV) and likely voter (LV) results. It turns out that the RV and LV results, showing a 49%-48% Kerry lead amongst LVs, and a 48%-48% tie amongst RVs, were once again based on samples that contained more Republicans than Democrats.
This is the information I obtained from Gallup this morning:
Likely Voter Sample Party ID – Poll of October 1-3
Reflected Dead Heat 49%-49%
Total Sample: 772
GOP: 220 (35%)*
Dem: 245 (39%)*
Ind: 175 (26%)*
*Repercentaged as party ID was not asked of every respondent and interviewers called back to ascertain this information from missed respondents.
Likely Voter Sample Party ID – Poll of October 9-10
Reflected Kerry Winning by 49%-48%
Total Sample: 793
GOP: 302 (38%)
Dem: 269 (34%)
Ind: 219 (28%)
Note that last week’s poll after the first debate had a historically more supportable 4% Democratic advantage in party ID, and Kerry had pulled into a tie with Bush amongst LVs. This, after trailing him allegedly by 8% amongst LVs with a sample that had a 12% GOP advantage before the first debate. Now Gallup is telling us that after the second debate, where Gallup’s own poll shows that Kerry won as well, Kerry has eked out a 1% lead over Bush amongst LVs, using a sample with a 4% GOP advantage.
So how big is Kerry’s lead if these results were recalibrated using a sample with a historically more supportable Democratic advantage?
Registered Voter Sample Party ID – Poll of October 1-3
Reflected Bush Winning by 49%-47%
Total Sample: 934
GOP: 274 (36%)*
Dem: 287 (38%)*
Ind: 209 (26%)*
*Repercentaged as party ID was not asked of every respondent and interviewers called back to ascertain this information.
Registered Voter Sample Party ID – Poll of October 9-10
Reflected Dead Heat 48%-48%
Total Sample: 1015
GOP: 364 (36%)
Dem: 344 (34%)
Ind: 295 (29%)
Similarly, Gallup’s RV results from yesterday show a tie, but with a sample that has moved from a 2% Democratic advantage back to a 2% GOP advantage in one week. Again, the question here is how big is Kerry’s lead amongst RVs now if the Gallup sample was recalibrated to reflect a Democratic edge in RVs?
So Kerry still manages to tie Bush in polls using samples with a GOP advantage. It does make you wonder how big a lead he would have if these samples eliminated the GOP bias.
More to the point: How big of an impact would it have on this race for the venerable Gallup to report that Kerry had assumed a 3-4% lead over Bush heading into the last debate?